Adaptive Moments are Surprisingly Effective for Plug-and-Play Diffusion Sampling
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914500089741312 |
|---|---|
| author | Belardi, Christian Lovelace, Justin Weinberger, Kilian Q. Gomes, Carla P. |
| author_facet | Belardi, Christian Lovelace, Justin Weinberger, Kilian Q. Gomes, Carla P. |
| contents | Guided diffusion sampling relies on approximating often intractable likelihood scores, which introduces significant noise into the sampling dynamics. We propose using adaptive moment estimation to stabilize these noisy likelihood scores during sampling. Despite its simplicity, our approach achieves state-of-the-art results on image restoration and class-conditional generation tasks, outperforming more complicated methods, which are often computationally more expensive. We provide empirical analysis of our method on both synthetic and real data, demonstrating that mitigating gradient noise through adaptive moments offers an effective way to improve alignment. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_16797 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Adaptive Moments are Surprisingly Effective for Plug-and-Play Diffusion Sampling Belardi, Christian Lovelace, Justin Weinberger, Kilian Q. Gomes, Carla P. Machine Learning Computer Vision and Pattern Recognition Guided diffusion sampling relies on approximating often intractable likelihood scores, which introduces significant noise into the sampling dynamics. We propose using adaptive moment estimation to stabilize these noisy likelihood scores during sampling. Despite its simplicity, our approach achieves state-of-the-art results on image restoration and class-conditional generation tasks, outperforming more complicated methods, which are often computationally more expensive. We provide empirical analysis of our method on both synthetic and real data, demonstrating that mitigating gradient noise through adaptive moments offers an effective way to improve alignment. |
| title | Adaptive Moments are Surprisingly Effective for Plug-and-Play Diffusion Sampling |
| topic | Machine Learning Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.16797 |